THE CAPABILITY OF THE MR DAMPER VERIFIED BY SHAKING TABLE TESTS AND REAL TIME HYBRID TESTS

Transcription

1 October 1-17, 8, Beijing, China THE CAPABILITY OF THE MR DAMPER VERIFIED BY SHAKING TABLE TESTS AND REAL TIME HYBRID TESTS H. Fujitani 1, H. Sakae, M. Ito, R. Kawasaki 3, A. Masutani 3, H. Fujii and T. Hiwatashi 4 ABSTRACT : 1 Professor, Dept. of Architecture, Kobe University, Kobe, Japan Former graduate student, Dept. of Architecture, Kobe University, Kobe, Japan 3 Graduate student, Dept. of Architecture, Kobe University, Kobe, Japan 4 Senior researcher, Research and Development Center, TOA Corporation, Yokohama, Japan (Former graduate student of Kobe University) fujitani@kobe-u.ac.jp Magnetorheological fluid damper (MR damper) has been expected to control the response of civil and building structures in recent years, because of its large force capacity and variable force characteristics. The important objective of this paper is to verify the validity of real-time s by comparison with the test results of shaking table tests by using the same MR damper. The maximum damping force of the MR damper is 1 (kn), the stroke is 6(p-p) (mm), and the maximum piston velocity is 1(m/s). To determine the control force of the MR damper, optimal control theory and skyhook control were employed. The capability of the MR damper to control the response displacements and accelerations of base isolation system was verified by both s and real-time s. And then, the capability of the MR damper to control the response displacements and accelerations of base isolation system was verified by sliding mode control and H control. KEYWORDS: Real time, Semi active control, Shaking table test, MR damper 1. INTRODUCTION Hybrid tests were developed, in order to make clear the behaviors of structures excited by earthquake ground motions. At first, pseudo-dynamic s were conducted. Then, the method of real-time s of passive structures was developed (Nakashima et al. 199). On the other hand, researches of semi-active control using Mgnetorheological fluid damper (MR damper) have been developed (for example, Rabinow et al. 1948, Yoshioka et al., Fujitani et al. 4). In the research of semi-active control there are three verification methods. One is the numerical simulation; the other one is the. And the last one is the real-time. In recent years, Christenson (Christenson et al. 8), Wu (Wu et al. 6), Carrion (Carrion et al. 7) and Iemura et al. worked on the real-time of semi-active control. Christenson et al. verify the accuracy of their method of real-time s by numerical simulation. Wu et al. verify by spectrum radius analysis method. Authors conducted s for base-isolation system and real-time s by using same MR damper, in order to verify the accuracy of the real-time (Fujitani et al. 8).. OBJECTIVES Authors have been worked on semi-active control of structures against earthquake ground motion, in order to improve the performance of the structure by using MR damper, not only to keep safety of human lives but also to keep function of buildings and comfort even in the duration of the ground motion (Fujitani et al. 4). To verify the effectiveness of semi-active control, authors conducted s. But the cost of shaking table tests is high, and many research institutes and university laboratories do not have shaking table. So, the system of real-time s of semi-active control is eagerly longed. Kobe University developed a real-time system, and the accuracy of the real-time system was verified by comparing results of the test with results of s in this paper.

2 October 1-17, 8, Beijing, China 3. TEST SPECIMEN 3.1. Magnetorheological (MR) damper The maximum damping force of the MR damper used in this study is 1kN and its stroke is +/-3mm as shown in Table 1. Figure shows the force-velocity relationship of the MR damper. The MR damper is modeled simply by the bingham plastic model as shown in Figure 3, and the relationship between the damper force and the electric current is approximated as Equation (1). Figure 4 shows the test results and simulated force-displacement relationship. The Bingham plastic model and Equation (1) shows good agreement between them. Reservoir Piston Bypass Flow Orifice Figure 1 Structure of MR damper Table 1 Dimensions of MR damper Maximum Force Stroke 1 kn ± 3 mm Electromagnet Max.Current 5. A Magnetorheological fluid Bando : 3 1 Force (kn) A 3A A 1A Velocity (cm/s) Figure Force-piston velocity relationship of MR damper P A C Figure 3 Bingham plastic model F MR = sign( v)( 19 I + 7 I + 611) v (1) Force (kn) Force (kn) Displacement (mm) Displacement (mm) (a) Test results (b) Simulated force-displacement relationship Figure 4 Comparison of simulated force-displacement relationship and test results

3 October 1-17, 8, Beijing, China 3.. Shaking table test specimen The s were conducted at Building Research Institute in Tsukuba, Japan. Figure 5 and Table show the test specimen of base isolation system with one degree of freedom system. The test specimen was used, when optimal control was adopted. Figure 6 and Table 3 show the test specimen of base isolation system with two degrees of freedom system. The test specimen was used, when sky-hook control was adopted. MR damper Mass Table Dimensions of testing system of SDOF Roller bearing Laminated rubber bearing Mass / Weight 1. ton Friction force of roller bearing.9 kn Stiffness of laminated rubber bearing 51.1 kn/m Damping coefficient.3 kn s/m Natural period 3.4 sec Figure 5 Testing system of base isolation (SDOF) MR damper Masses Roller bearing Laminated rubber bearing Figure 6 Testing system of base isolation (Two DOF) Table 3 Dimensions of testing system (DOF) nd floor Mass / Weight 8.1 ton Stiffness of lamina te d rubber bea ring 3.5 kn/cm Da mping coe fficient.3 kn s/cm Na tural period.95 sec Base-isola te d layer Mass / Weight 6.3 ton Stiffness of lamina te d rubber bea ring.5 kn/cm Da mping coe fficient.3 kn s/cm Friction force of roller bearing.9 kn Na tural period 3.43 sec 4. REALTIME HYBRID TEST A high velocity test machine with feed back control by DSP was set up in Kobe University (Figure 7). The maximum velocity is 1 (m/s), and the maximum force is 1 (kn) and the stroke is 6(p-p) (mm). The damper force is measured and input to DSP every 1/1 or 1/5 seconds. The response displacement and velocity are calculated by solving the equation of motion (differential equation) of the structure including the damper force, and the control variable (DC electric current, in case of MR damper) is calculated, too. The calculated response displacement and velocity are input to the test machine. The calculated control variable (DC electric current) is induced to the MR damper by electric power supply. 5. COMPARISON OF SHAKING TABLE TESTS AND REALTIME HYBRID TEST Optimal control theory (Yang 1975) and skyhook control (Karnopp 1974) were adopted.

4 October 1-17, 8, Beijing, China DC Electric current is induced. Damper force Load cell MR damper Slide table AC servomotor Control Calculation DSP Response Calculation Response displacement and velocity of base isolation are input. Figure 7 Real-time system in Kobe University 5.1 Optimal control In the case of the optimal control theory, authors apply the evaluation function shown in Equation (). 1 J = [ α d x + α v x& + α a (&& x + & z ) + γu ]dt () α d,α v,α a,γ are weighting coefficients, related to displacement, velocity, acceleration, control force. In this paper, α d,α v,α a were fixed to 1,1,1 and γ was fixed to.1. The absolute acceleration, the relative displacement and the control force are shown in Figures Against two kinds of earthquake ground motions (El Centro 194 NS, JMA Kobe 1995 NS), acceleration, displacement and control force among, hybrid test and simulation agree very well. 5. Skyhook control The control force is generated by skyhook control theory, when the product of the absolute velocity and relative velocity is positive. The control force F(t) is decided by Equation (3). F t = λ f t λ = 1 (3) () ( ) ( ) 1 1 f () t = mu& + ku (4) t t f : sum of kinetic energy and strain energy (kj), m : mass of superstructure (t), () t k : stiffness of base isolation layer (kn/m), u t, u& t : displacement (m) and velocity (m/s) of superstructure The absolute accelerations of each mass, the relative displacements of each story, the induced electric current and the control force are shown in Figures 9. Against the earthquake ground motions of El Centro 194 NS. Acceleration of each story, displacement of each story, electric current and control force among shaking table test, and simulation agree very well. 6. APPLICATION OF REALTIME HYBRID TEST 6.1 Sliding mode control (SMC) The state equation and switching function of a structure are represented by Equation (5). X& = AX + Bu (5) σ = SX = First S which represents the inclination of the hyper plane is selected as the feedback gain of optimum control

5 October 1-17, 8, Beijing, China as shown in Equation (6). 1 T S = R B P (6) P is a solution of the Riccati equation as shown in Equation (7) T 1 T PAε + Aε P PBR B P + Q = (7) Aε = A + εi( ε ) For the switching control rule, the eventual sliding mode control scheme is used. Controlling force u is given by Equation (8). acceleratio n (mm/s ) displacement (mm) force (kn) Figure 8.1 Comparison of, and numerical simulation in case of El Centro 194 NS acceleratio n (mm/s ) simulation displacement (mm) force (kn) simulation Figure 8. Comparison of, and numerical simulation in case of JMA Kobe 1995 NS

6 October 1-17, 8, Beijing, China 1 σ u = ( SB) SAX η (8) σ + γ Lyapunov function V is selected as shown in Equation (9), and η is obtained so that its first-order differential becomes negative. V =.5σ (9) V & < acceleratio n (mm/s ) F_Acc (mm/s ) simulation displacement (mm) F_Disp (mm) acceleratio n (mm/s ) F_Acc (mm/s ) displacement (mm) F_Disp (mm) simulation shaking table tes t force (kn) Force (kn) simulation current (A) Current (A) Figure 9 Comparison of, and numerical simulation in case of El Centro 194 NS

7 October 1-17, 8, Beijing, China SMC and some control methods were applied to single-degree-of-freedom system shown in Figure 1 and Table 4; a comparative study of the response suppression performances of SMC was performed. Comparison of the absolute acceleration and response displacement at SMC and at non-control state, time-history of controlling force of SMC are shown in Figure 11. El Centro 194 NS was used as the input earthquake ground motion. The test results of following cases were compared. Case 1) No damper Case ) Passive control when constant electric current is applied (,.5, 1, 1.5,, 3, 4 (A)) Case 3) Passive control by constant damping coefficient (1,, 3, 5, 1 (kns/m)) Case 4) Semi-active control by optimal control theory Case 5) Semi-active control by Sliding Mode Control Figure 11 shows the relationship between the maximum displacement and maximum absolute acceleration. Both the maximum displacement and the maximum absolute acceleration were reduced to about one-fourth by SMC compared to those uncontrolled. Furthermore, the maximum displacement was nearly equal to the minimum value by another control method; the maximum acceleration was about one-half, which revealed that SMC reduced both displacement and acceleration. Maximum 最大加速度 acceleration [m/s ] Large dampin 減衰大 Small 減衰小 最大変位 [mm] Maximum 最大加速度 acceleration [m/s ] 最大変位 [mm] Maximum displacement Maximum displacement (a) Numerical simulation (b) Real time Figure 11 Relationship of maximum displacement and maximum acceleration Large 減衰大 dampin Small 減衰小 Figure 1 Single degree of freedom system Table 4 Characteristics of the system Mass (kg) 6 Stiffness (kn/m) 4 Damping factor. Natural period (s).994 :without damper :constant electric current :constant damping coefficient :optimal control :sliding contol mode 6. H Control The real-time system was applied to a single degree of freedom system by using H Control. Although the control system designed to the natural period of T=1. (s), the actual natural period Ta may be.8 (s) or 1. (s). The results are shown in Table 5 (acceleration) and Table 6 (displacement). In case of Ta=.8, the reduction ratio became worse a little than the case of Ta=1.. On the other hand, in case of Ta=1., the reduction ratio was almost same as case of Ta= CONCLUSIONS A real-time system was developed for the test of semi-active control in Kobe University, and the accuracy of the system was verified by comparing results of the test with results of s in this paper. Base-isolated structure was controlled by optimal control theory. And Application by using sliding mode control and H Control were introduced. ACKNOWLEDGEMENT The authors wish to thank to Dr. Taiki Saito and Mr. Namihiko Inoue of Building Research Institute.

8 October 1-17, 8, Beijing, China Table 5 Comparison of the maximum absolute acceleration of real-time and simulation El Centro-NS max. absolute acceleration simulation real time Ta=.8 (designed T=1.) A H control A H control max. absolute acceleration (mm/s) reduction ratio (%) accelaration RMS value (mm/s) Ta=1. (designed T=1.) A H control A H control max. absolute acceleration (mm/s) reduction ratio (%) accelaration RMS value (mm/s) Ta=1. (designed T=1.) A H control A H control max. absolute acceleration (mm/s) reduction ratio (%) accelaration RMS value (mm/s) Table 6 Comparison of the maximum displacement of real-time and simulation El Centro-NS max. displacement simulation real time Ta=.8 (designed T=1.) A H control A H control max. displacement (mm) reduction ratio (%) Ta=1. (designed T=1.) A H control A H control max. displacement (mm) reductiion ratio (%) Ta=1. (designed T=1.) A H control A H control max. displacement (mm) reduction ratio (%) REFERENCES Nakashima, M., Kato, H. and Takaoka, E. (199). DEVELOPMENT OF REAL-TIME PSEUDO DYNAMIC TESTING, Earthquake Engineering and Structural Dynamics, Vol.1, 79-9 Rabinow, J. (1948). The Magnetic Fluid Clutch, AIEE Trans., Vol.67, 13. Yoshioka, H., J.C. Ramallo, and B.F. Spencer, Jr. (). Smart Base Isolation Strategies Employing Magnetorheological Dampers, Journal of Engineering Mechanics, ASCE, Vol. 18, Fujitani, H., H. Sodeyama, K. Hata, T. Hiwatashi, Y. Shiozaki, N. Inoue and S. Soda. (4). Application of Magnetorheological Fluid to Semi-Active Control of Building Structures by BRI and Partners, Key Engineering Materials Vols. 7-73, Christenson, R., Lin, Y., Emmons, A., and Bass, B. (8). Large-Scale Experimental Verification of Semiactive Control through real-time Hybrid Simulation, Journal of Structural Engineering, ASCE, Vol.134, No.4, pp Wu, B., Wang, Q., Benson Shing, P. and Ou, J. (7). Equivalent force control method for generalized real-time substructure testing with implicit integration, Earthquake Engineering and Structural Dynamics, Vol.36, Carrion, Juan E. and Spencer, Jr., Billie F. (7). Model-based Strategies for Real-time Hybrid Testing, Newmark Structural Engineering Laboratory Report Series 6, University of Illinois at Urbana-Champaign. Yang, J.N.. (1975). Application of Optimal Control Theory to Civil Engineering Structures, Journal of The Engineering Mechanics Division, pp Karnopp, D., Crosby, M. J. and Harwood, R. A.. (1941). Vibration Control Using Semi-Active Force Generators, Journal of Engineering for Industry, Transactions of ASME, Hideo Fujitani, Hiroaki Sakae, Rui Kawasaki, Hideki Fujii, Takeshi Hiwatashi and Taiki Saito. (8). Verification of real-time s of response control of base isolation system by MR damper comparing s, Proceedings of the SPIE's 15th Annual International Symposium on Smart Structures and Materials, No

9 October 1-17, 8, Beijing, China

10 October 1-17, 8, Beijing, China

11 October 1-17, 8, Beijing, China